MODEL DIRECTORY

Latest AI models

New releases, in one place. Explore models by their publication date.

Newest first7 languages
Clear

562 models

Newest first
Zhipu AI

GLM-4-32B-Base-0414

GLM-4-32B-Base-0414 is a base language model with approximately 32 billion parameters in the GLM-4-0414 series, intended for text generation, further training, and local deployment. The official…

Zhipu AI

GLM-Z1-Flash

GLM-Z1-Flash is a free reasoning model in the GLM-Z1 series, intended to lower the barrier to using reasoning models. The official release record is dated 2025-04-14. (docs.bigmodel.cn) Deep thinking…

Zhipu AI

GLM-Z1-Air

GLM-Z1-Air is a cost-effective deep-thinking reasoning model launched by Zhipu AI for high-frequency call scenarios. It was officially announced on April 14, 2025; the version identifier for this…

Zhipu AI

GLM-Z1-AirX

GLM-Z1-AirX is the high-speed version of GLM-Z1-Air, positioned as an ultra-fast reasoning model. It was officially released on April 14, 2025, with the release identifier GLM-Z1-AirX-0414.…

Zhipu AI

GLM-4-Flash-250414

A free text foundation model in the GLM-4 series, intended for general-purpose tasks and lightweight applications. Zhipu released this version on April 14, 2025. (docs.bigmodel.cn) Takes text as input…

Zhipu AI

GLM-4-Air-250414

GLM-4-Air-250414 is a cost-effective base language model released by Zhipu on April 14, 2025, for complex tasks and agent applications. (docs.bigmodel.cn) The model accepts text input and generates…

Google

Chirp 3: Transcription

Chirp 3: Transcription is a Google generative model for multilingual automatic speech recognition (ASR) that transcribes audio into text. It is available through the Google Cloud Speech-to-Text API V2…

Moonshot AI

Kimi-VL-A3B-Thinking

Kimi-VL-A3B-Thinking is a vision-language reasoning model in the Kimi-VL series, focused on complex text and multimodal reasoning. It uses a mixture-of-experts (MoE) architecture, with approximately 1…

Moonshot AI

Kimi-VL-A3B-Instruct

Kimi-VL-A3B-Instruct is a vision-language instruction model designed for efficient inference. It uses a mixture-of-experts (MoE) language model, the native-resolution vision encoder MoonViT, and a…

Alibaba

Wan 2.1 KF2V Plus

Wan 2.1 First-and-Last-Frame-to-Video Plus is a model for international video generation services: it generates a transition video from two images representing the first and last frames, and can also…

MiniMax

Speech-02-Turbo

Speech-02-Turbo is a text-to-speech model in the MiniMax Speech-02 series, optimized for real-time interaction and focused on balancing low latency with voice quality. The Speech-02 series was…

MiniMax

Speech-02-HD

Speech-02-HD is a text-to-speech model in MiniMax's Speech-02 series designed for high-fidelity speech generation, suitable for applications that prioritize audio quality, such as dubbing and…

Alibaba

QVQ-Max

QVQ-Max is a visual reasoning model in the Tongyi Qianwen QVQ series, designed for visual understanding and deep-thinking tasks. Alibaba Cloud Bailian's model release record lists qvq-max on March 26,…

Alibaba

Wanx2.1-ImageEdit

Wanx2.1-ImageEdit is a Wanxiang Tongyi image editing model for local or full-image editing of existing images. It supports preset editing tasks and instruction-based editing, with both input and…

Tencent

hunyuan-turbos-longtext-128k-20250325

The long-text version of Tencent Hunyuan TurboS, further trained on the TurboS base model and using a mixture-of-experts (MoE) architecture. The release date recorded in Tencent Cloud product updates…

Tencent

Hunyuan-T1

Hunyuan-T1 is a deep reasoning large model released by Tencent. This version was released on March 21, 2025; the official T1 release is based on the Hunyuan TurboS foundation model, which uses a…

Alibaba

GTE-Rerank-V2

GTE-Rerank-V2 is a multilingual text reranking model developed by Tongyi Lab. It reorders documents based on their semantic relevance to a query and is suitable for scenarios such as semantic search…

Alibaba

Qwen-Plus-Character

Qwen-Plus-Character is a text generation and role-playing model in the Qwen series, designed for anthropomorphic dialogue. It is a dynamically updated version, and the official documentation states…

Baidu

ERNIE X1

ERNIE X1 is a deep-thinking reasoning model released by Baidu. Baidu says it has multimodal capabilities and enhanced understanding, planning, and reflection capabilities. (prnewswire.com) The model…

Baidu

ERNIE 4.5

A preview model in Baidu's ERNIE 4.5 series. Baidu positions ERNIE 4.5 as a natively multimodal foundation model and provides Qianfan API services to developers and enterprises.…